a comparative study of the interaction of 5,10,15,20-tetrakis (n-methylpyridinium-4-yl)porphyrin and...

volume 13 Number 1 1985 Nucleic Acids Research

A comparative study of the interaction of 5,10,15,20-tetrakis (N-methylpyridinium-4-yl)porphyrinand its zinc complex with DNA using fluorescence spectroscopy and topoisomerisation

John M.Kelly and Martin J.MurphyDepartment of Chemistry, University of Dublin, Trinity College, Dublin 2, Ireland, and

David J.McConnell and Colm OhUiginDepartment of Genetics, University of Dublin, Trinity College, Dublin 2, Ireland

Received 26 October 1984; Revised and Accepted 7 December 1984

ABSTRACT

Binding of 5,10,15,20-tetrakis(N-methylpyridinium-4-yl)porphyrin(l̂ TMPyP̂ "1") and its zinc complex (ZnTMPyP^+) to DNA is demonstrated by theircoelectrophoresis and by absorption and fluorescence spectroscopic methods.Topoisomerisation of pBR322 DNA shows that H2TMPyP

4+ unwinds DNA as effic-iently as ethidium bromide showing that it intercalates at many sites.ZnTMPyPz>+ may cause limited unwinding. Marked changes in the fluorescencespectra of the porphyrins are found in the presence of DNA. The fluorescenceintensity of either H2TMPyP4+ Or ZnTMPyP^ is enhanced in the presence ofpoly (d(A-T)), whereas in the presence of poly (d(G-C)) the fluorescenceintensity of ZnTMPyP^ is only slightly affected and that of H2TMPyP4+markedly reduced. Both the porphyrins photosensitise the cleavage of DNAin aerated solution upon visible light irradiation.

INTRODUCTION

Although meso-substituted porphyrins are much larger than well-known

intercalators such as acridine or phenanthridine dyes, it is now apparent

from the work of Fiel, Pasternack and their coworkers (1-11) that certain

cationic water-soluble porphyrins can interact strongly with DNA in solution.

It is clear that there is more than one mode of binding and the nature of

this interaction depends on the structure of the porphyrin, on the relative

concentrations of the porphyrin and the DNA base pairs, and on the base

composition of the DNA. Some porphyrins appear to be able to intercalate.

The best evidence for this is the effect of the porphyrin on the electro-

phoretic mobility of DNA (2) and on the viscosity of DNA solutions (1,9)

and from circular dichroism studies (1,6,7,8,10). Porphyrins which do not

intercalate are presumed to bind to the outside of the DNA, probably

interacting both with the bases and the phosphate groups (11). Generally

it has been noted that porphyrins which intercalate do not have axial

ligands and have meso-groups which can easily rotate into the plane of the

tetrapyrrole ring (8,10,11).

The base specificity of this interaction has been studied by both abs-

© IRL Press Limited, Oxford, England. 167

at Laurentian U

niversity on Novem

ber 5, 2014http://nar.oxfordjournals.org/

Dow

nloaded from

http://nar.oxfordjournals.org/

Nucleic Acids Research

Figure 1. Structural formulae of5,10,15,20-tetrakis(N-methylpyrid-inium-4-yl)porphyrin and its zinccomplex.

orption spectroscopy (10) and circular dichroism (8,10). Porphyrins which

are believed to intercalate (e.g. HjTMPyP^1* - see Figure 1 - or its CU(II)

or Ni(II) derivative) appear to interact strongly with poly(d(G-C)). In

contrast the spectroscopic properties of non-intercalating porphyrins are

hardly changed by poly (d(G-C)) but are markedly altered by poly (d (A-T)).

This suggests that intercalation occurs predominantly at residues. The

kinetics of association and dissociation of porphyrins and DNA or synthetic

polynucleotides has been treated in detail by Pasternack et al. in a recent

paper (11).

We have recently started to study how the photoproperties, including the

non-linear optical properties, of organic molecules and metal complexes are

affected by binding to DNA. The known photochemical and photophysical

properties of porphyrins (12) make them particularly attractive as probes

for the behaviour of DNA and its complexes in solution and also as sensitis-

ers for the photocleavage (possibly base-specific) of polynucleotides.

Initially we have chosen to study two water-soluble porphyrins (H2TMPyP^+

and ZnrMPyP^+) which fluoresce strongly in solution. As indicated above,

evidence suggests that H2TM?yP^+ intercalates into DNA, but ZnTMPyP^+ does

not. We have investigated this matter by carrying out a sensitive assay

for intercalation, namely the topoisomerisation of supercoiled covalently

closed circular (ccc) DNA in the presence of the porphyrins. We demonstrate

that the fluorescence spectrum and intensity of the porphyrins are changed

substantially in the presence of DNA and that the photophysical properties

of the porphyrins are sensitive to the base composition of the polynucleo-

tide. It is shown that both porphyrins are equally effective in causing

photo-cleavage of the polynucleotide in aerated solutions (4).

168

at Laurentian U

niversity on Novem


Dow

nloaded from



MATERIALS AND METHODS

Porphyrins. 5,10,15>20-tetrakis(4-pyridyl)porphyrin (H2TPyP) was

prepared by refluxing 4-pyridylcarboxaldehyde and pyrrole in a 70:1 propionic

acid/acetic acid mixture (13). The purified H2TPyP was then refluxed with

methyl-p-toluene sulphonate in dimethyl formamide (DMF) to yield H~TMPyP

as the tetratosylate salt (14). Extinction coefficient at 423 nm (2.2 x 10 )

is in agreement with the literature (10). ZnTMPyP . 4C1*was obtained in4+

three ways: (i) by refluxing H.TMPyP with a 300-fold excess of ZnCl. in

DMF (15,16), (ii) by refluxing H2TMPyP + with a 500-fold excess of ZnCl

in water (10), and (iii) commercially manufactured by Midcentury Ltd. The4+

absorption spectra for ZnTMPyP solutions from the above three sources

were identical and the extinction coefficient at 437 nm was taken to be

2.3 x 105 (17,18).

DNA and polynucleotides. The plasmid pBR322 was isolated from

E.coli K-12 C600 by standard procedures (19), using chloramphenicol

amplification. The plasmid DNA was further purified on CsCl/ethidium

bromide density gradients. The ccc DNA was extracted by isopropanol four

times, dialysed against TE buffer (10 mM Tris, 1 mM EDTA, pH 7.4), and

stored at 4 C. High molecular weight calf thymus DNA was obtained from

Sigma (Cat. No. D4764). It was dissolved in water (100 mg in 50 ml of

doubly distilled water), made to 1% sodium dodecyl sulphate (SDS) and 1 M

NaCl before extraction with an equal volume of a phenol:chloroform:isoamyl

alcohol mixture (25:24:1) containing 0.5% w/v of 8-hydroxyquinoline. This

solvent mixture had been saturated by standing over aqueous 100 mM Tris

HC1 buffer pH 7.5. This extraction was repeated once, and the DNA

spooled out on a glass rod while adding 2 volumes of ice cold ethanol.

The DNA was dissolved in TE buffer to a final concentration of approximat-

ely 2.8 x 10 M of phosphate and dialysed extensively against TE buffer.

Storage was over a few drops of chloroform at 4 C. Polynucleotides

poly[d(G-C)] and poly[d(A-T)] were obtained from PL Biochemicals (Cat. No.

7910 and 7870 respectively).

Co-electrophoresis of DNA-porphyrin complexes: DNA-porphyrin com-

plexes were prepared in 10 uH reaction mixtures containing 0.2 yg of

pBR322 DNA and porphyrin, as specified, in TBE buffer (0.089 M Tris base,

0.089 M boric acid, 0.002 M EDTA). The reaction mixtures were allowed

to stand for 10 minutes at room temperature before being electrophoresed

for 3 hours in a horizontal agarose (0.9% w/v) gel (20 cm x 10 cm x 0.4 cm)

at a constant 200 volts. The reservoir and gel buffers were TBE. The

169

at Laurentian U

niversity on Novem


Dow

nloaded from



gel was photographed using transmitted light to detect the porphyrin. It

was then stained for 30 minutes in ethidium bromide (0.5 ug/ml) before

photographing using a long wave ultra violet light source to induce

fluorescence of the ethidium bromide. Finally the gel was allowed

to stand for 48 hours at room temperature in ethidium bromide solution be-

fore photography as before. This last procedure was found necessary to

allow porphyrin bound to DNA to dissociate and ethidium bromide to bind.

Topoisomerisation of DNA and DNA-porphyrin complexes. The enzyme

topoisomerase I (Calf thymus, Bethesda Research Laboratories; Cat. No.

8042 SA/SB) was used to convert supercoiled ccc pBR322 DNA to topoisomers,

by the procedure of Keller (20). Samples of pBR322 DNA (in excess of 80%

ccc DNA) containing ca. 1.6ug DNA in 50ul of the recommended reaction

buffer (50mM Tris, 50mM KC1, lOmM MgCl2, 0.5mM DDT, O.lmM EDTA, 30yg/ml

bovine serum albumin), topoisomerase and porphyrin or ethidium at the speci-

fied concentration (see Figure 3) were incubated at 37 C for 4 hours. The

reaction was stopped by the addition of an equal volume of the phenol:

chloroform:isoamyl alcohol :8-hydroxyquinoline mixture and the aqueous layer

recovered after shaking and centrifugation. This extraction is known to

remove ethidium bromide and also removes the porphyrins used in these ex-

periments. The DNA was recovered by ethanol precipitation after the aqueous

solution had been made 0.3M in sodium acetate. After drying, the precipi-

tate was redissolved in lOul of TBE buffer and electrophoresed on submerged

horizontal agarose gels (0.9% w/v, measuring 20cm by 22cm by 0.4cm) for 12

hours on TBE buffer at room temperature. The gel was stained for 30

minutes in ethidium bromide (0.5ug/ml) and excess ethidium bromide was re-

moved by standing in magnesium sulphate (lmM) for 20 minutes before photo-

graphing under ultra violet light as before. e at 460 nm for ethidium

bromide was taken to be 4220 (20).

Absorption and Fluorescence Spectra. Absorption spectra were obtained

either on a Pye-Unicam SP82OO UV/visible spectrometer or on a Cary 219 IIV/

visible spectrometer interfaced to a PDF 11/34 minicomputer for data analy-

sis. Fluorescence measurements were recorded on a Perkin Elmer MPF-44B

fluorimeter and the spectra were not corrected for the wavelength response

of the (R9?8) photomultiplier. The excitation wavelengths were 423 nm for

all H2TMPyP solutions, and 437nm for all ZnTMPyP + solutions (unless

otherwise stated). Fluorescence spectra were carried out in a 10mm path-

length quartz cuvette at room temperature in a phosphate buffer containing

O.O25M K2HP04and O.O25M KH2PO4 (pH = 6.9). Optical densities at the excitat-

170

at Laurentian U

niversity on Novem


Dow

nloaded from



4+ 4+

ion wavelength were 0.5 or less. As H.TMPyP and ZnTMPyP tend to stick to

surfaces including glass (21), procedures were used to minimise the effect of

this on quantitative fluorescence or absorption spectra. For example, for

fluorescence measurements the clean cuvette was first flushed with a portion

of the prophyrin sample and then with distilled water. A further aliquot

(1.0ml) of the porphyrin was added to the cuvette, shaken, and the spectrum

taken. The required amount of DNA was then added to this and the spectrum

taken. Sodium chloride, when required, was added as solid. Porphyrin-

sticking is also reduced by treatment of glassware with Me_SiCl . Results

obtained by such procedures were reproducible.

Photolysis: Solutions of DNA and porphyrin were prepared in the dark

by adding 25y£ of a lOyM porphyrin solution to 235yit of a solution contain-

ing 4yg of pBR322 DNA in a 6mm diameter glass tube. Sodium azide (lOOmM)

was present when specified. The solution was mixed gently by pipetting

and irradiated with 436 nm light isolated from a medium pressure mercury

lamp(Thorn MED 25OW) using a high radiance monochromator. Sample tubes

were clamped in pairs 3cm distant from the monochromator, aliquots (40]i£)

were removed at specified times and added to 50IJ£ of phenol (buffered by

storage over 1M Tris HC1 pH 7.5) before vigorous mixing to extract the

porphyrin from the aqueous phase. The aqueous phase was reduced by lyo-

philisation to about 20u£ for electrophoresis over-night on a agarose gel

(0.8 w/v) at 4 volts per cm. The gels were stained with ethidium bromide

and photographed as before.

RESULTS4+

Co-electrophoresis of DNA and porphyrins. The porphyrins H TMPyP

and ZnTMPyP^ , being positively charged, migrate to the cathode when subjec-

ted to electrophoresis in TBE buffer at pH 8.3. DNA, being negatively

charged, moves in the opposite direction. If these porphyrins bind strongly

to DNA they are expected to co-electrophorese with the DNA, and under the

conditions of low ionic strength they should reduce the electrophoretic4+

mobility of the DNA. These effects were observed with both H.TMPyP and4+

ZnTMPyP . Solutions containing pBR322 DNA were prepared with increasing

concentrations of porphyrins (OuM, 3uM, 7uM, 50yM) and electrophoresed.

Porphyrins when alone migrated towards the cathode as a smear and this was

observed easily from its yellow-green colour. When DNA was mixed with a low

concentration of porphyrin, no porphyrin was observed moving to the cathode,

but instead two narrow yellow-green bands were observed moving to the anode

171

at Laurentian U

niversity on Novem


Dow

nloaded from



corresponding to ccc DNA and oc DNA. If the concentration of porphyrin was

increased to 50uM, (P/D = 1.2) three coloured bands were observed, one a

smear towards the cathode typical of free porphyrin and the other two towards

the anode apparently corresponding to ccc and oc DNA. The positions of the

ccc and oc DNA were confirmed by staining with ethidium bromide. After

staining the gel for 30 minutes in ethidium bromide, DNA complexed with a

high concentration of porphyrin was not detected or showed only as faint

bands, but after staining for 48 hours these bands were easily detected.

This indicated that either the binding of ethidium bromide to DNA is reduced4+ 4+

if the DNA is already complexed with HjTMPyP or ZnTMPyP , or that the

fluorescence of ethidium bromide-DNA complexes is quenched by the porphyrins.

The gels showed that the DNA-porphyrin complexes which are formed at higher4+ 4+

concentrations of H.TMPyP and ZnTMPyP have a lower electrophoretic

mobility as expected for complexes between the polyanionic DNA and the cat-

ionic porphyrins.

Topoisomerization of pBR322 and pBR322-porphyrin complexes. The

enzyme topoisomerase I catalyses the nicking and closing of both positively

and negatively supercoiled ccc DNA molecules, producing a series of topoiso-

mers, i.e. DNA molecules which differ in winding number and therefore in

the number of supercoils. A population of topoisomers shows a nearly normal

distribution of molecules which differ in the number of superhelical turns

by unit amounts distributed about a mean which is affected by the conditions

of the reaction - temperature, salt concentration etc. - at the time of ring

closure by topoisomerase I. Keller showed that agents which were known to

unwind DNA, such as ethidium bromide, could be studied through their effect

on the distribution of topoisomers formed by topoisomerase I (20). His

method provides an elegant way of demonstrating unwinding and quantifying

the efficiency of the effect of an unwinding agent. We have used this method4+ 4+

to determine whether H-TMPyP and ZnTMPyP unwind DNA,using the known inter-

calator ethidium bromide as a standard.4+

The addition of increasing quantities of the porphyrin lUTMPyP

to the topoisomerization reaction shifts the centre of the distribution of

topoisomers towards that of native negatively-supercoiled cccDNA (Figure 2).

Note that the temperature and ionic strength conditions of electrophoresis

and topoisomerisation differ so that the distribution of topoisomers formed

in the absence of unwinding agent is centred at ca. +3 supercoils. Addition

of unwinding agent, either H.TMPyP or ethidium bromide brings the centre of

the distribution initially through 0 supercoils. This explains the apparent

172

at Laurentian U

niversity on Novem


Dow

nloaded from



4 2 3 4 5 6 7 8 9101112131415 16

Figure 2. Unwinding of DNA by H?TMPyp4+. Samples of pBR322 plasmid DNA weretreated with topoisomerase I in the presence and absence of l̂ TMPyP̂ "1" andZnTMPyP^+ as described in Materials and Methods. The reactions were stopped,the samples phenol-extracted and electrophoresed. pBR322 DNA not treatedwith topoisomerase I consists of oc and ccc DNA only (lanes 2 and 10) pBR322DNA cut with the enzyme Hindlll, for which there is a single site on theplasmid, shows one band of DNA only (lanes 1 and 9). pBR322 DNA treatedwith topoisomerase I in the absence of either porphyrin (lanes 3 and 11)shows bands corresponding to topoisomers with a mean of +3 supercoils.pBR322 DNA treated with topoisomerase I in the presence of increasing amountsof H2TMPyP^+ (lanes 4 to 8) shows topoisomer bands with increasing numbers ofnegative supercoils. ZnTMPyP^+ (lanes 12-16) has a much smaller effectthan H2TMPyP^

+.

decrease in the mean of the distribution at low concentrations of unwinding

agent. The positively supercoiled molecules (lanes 3,4 and 11) migrate more

slowly than negatively supercoiled molecules (lanes 5 to 8). The effects on

the distributions caused by H.TMPyP and ethidium bromide were quantified

by estimating the means (centres) of the distributions and plotting the

change in the means against the concentration of unwinding agent (Figure 3).

The least squares lines give the change in the number of supercoils per uM

concentration of unwinding agent of 5 - 1 for ethidium bromide (c.f. 4.9 in

reference 20) and 6 - 1 for H TMPyP +. ZnTMPyP (lanes 12 to 16) has a

much smaller effect. It causes a maximum negative change of 2 - 3 supercoils.4+

The limit is not due to inhibition of the topoisomerase. ZnTMPyP added at

a P/D = 20 does not inhibit the effect of ethidium bromide (P/D = 20) on4+

topoisomerisation. The limit suggests that ZnTMPyP unwinds DNA through

interactions at a small number of sites on the DNA. Alternatively it is

possible the effect of ZnTMPyP is due to a derivative of ZnTMPyP formed

1*3

at Laurentian U

niversity on Novem


Dow

nloaded from



ETHIDIUH BROniDE -O

Figure 3. Change in supercoilinginduced by H2TMPyP't+ and ethidiumbromide. Change in the mean num-ber of supercoils at the centreof the topoisomer distributionplotted against concentration ofintercalator required to inducethat change. The lines throughthe points are derived from aleast squares analysis. pBR322at a concentration of 105 uMnucleotide.

0.8O 1.60 2.40 3.20

DYE COMCENTRflTIOM CHn

by reaction with a limited trace amount of impurity in the topoisomerisation

reaction.

Absorption and Fluorescence Spectra. The effect on the visible absor-

ption spectra of adding a large excess of CT-DNA to a solution of either4+ 4+

HTMPyP or ZnTMPyP in 0.05M (pH = 6.9) potassium phosphate buffer with no

added sodium chloride is shown in Figure 4. In close agreement with the4+

data of Pasternack et al. (10), we find that for H2TMPyP there is a sub-

stantial shift in the Soret band peak position (from 423 to 437 nm) and a4+

pronounced hypochromicity. In contrast, for ZnTMPyP there is only a muchsmaller shift (from 437 to 439 nm) and the extinction coefficient is only

slightly affected. The wavelengths of the Q-bands of the porphyrins are4+

also altered by DNA-interaction, those for H.TMPyP being red-shifted,4+

while those for ZnTMPyP move to the blue. When sodium chloride is added

to these porphyrin~DNA solutions (to a final concentration of NaCl of 1 M)

all these shifts are largely reversed (Figure 4b).

The influence of DNA base composition on these spectral shifts was exam-

ined by studying the synthetic polynucleotides poly(d(G_C)) and poly(d(A-T)).4+

For H^TMPyP the peak position of most bands are substantially red-shifted

for poly (d-(G-C)) but less so for poly(d(A-T)). The Soret band at 423 nm

shifts to 445 nm in the presence of poly(d(G-C)) and to 431 nm with poly4+

(d(A-T)) in u = 0.2 M buffer. On the other hand, for ZnTMPyP there are

only small red shifts in the Soret band position for either poly(d(G-C)) or

poly(d(A-T)) and a small blue shift of the Q-bands with poly(d(A-T)). As

expected CT-DNA (42% GC) shows intermediate behaviour.

174

at Laurentian U

niversity on Novem


Dow

nloaded from



/i

\\\\\,

M

1

A

H U H

100 ISO

\

i / \

1

11

//

i

M\

;

B- 4 —

1 i

1

— •

i - -

A •ift

_)11 ;

1 ,

, j

c

ri |

V > ^ _ _ I ,SM MO (SO

Figure 4. Absorption spectra ofporphyrin - CT-DNA complexes

(A) 2.2 pM H2TMPyP4+

2.0 uM H2TMPyP4+ and 166

UM CT-DNA (P/D = 82)

(B) 2.0 yM H2TMPyPA+ and 166

yM CT-DNA (P/D = 82)as above + 1 M NaCl.

(C) 2.2 yM ZnTMPyP'4+

2.1 yM ZnTMPyP^+ and 131yM CT-DNA (P/D = 64)

All solutions in 0.05 M phosphatebuffer.

The fluorescence spectra of each of the porphyrins were taken in the

presence of a large excess of CT-DNA, poly[d(G-C)] and poly[d(A-T)], (see

figures 5 and 6); y = 0.2 M buffer solutions and identical excitation con-

ditions (423 run for H-TMPyP and 437 nm for ZnTMPyP )were used for each

series. For H.TMPyp4+, the interaction with CT-DNA causes splitting of

the broad fluorescence band (centred at about 675 nm in the absence of DNA)

into two distinct bands with peaks at 654 and 714 nm. (A further very

weak band may also be discerned at about 615 nm. This feature is more pro-4+

nounced if solutions of H.TMPyP were not freshly prepared and might be

caused in part by an impurity). As is also shown in Figure 5, the base

composition of the polynucleotide exerts a strong influence on the emission

intensity from H-TMPyP . With poly[d(A-T)] present, two bands are noted

at 655 and 715 nm, the intensity being greatly enhanced compared to that of

H.TMPyP alone. By contrast, in the presence of poly[d(G-C)] the emiss-

175

at Laurentian U

niversity on Novem


Dow

nloaded from



W A V E L E N G T H nm.

Figure 5. Fluorescence spectraof H?TMPyP '̂r-DNA complexes^

(A) 2.3 uM H2TMPyP4+

2.1 pM H2TMPyP4+ and 420uM CT-DNA (P/D = 200)••• as above + 0.86 M NaCl.

(B) 3.2 pM H2TMPyP4+

2.9 uM H2TMPyP4+ and 110

uM poly[d(G-C)] (P/D = 38).••• as above + 0.86 M NaCl.

(C) 3.2 pM H TMPyPA+

2.9 pM H2TMPyPA+ and 140

uM poly[d(A-T)] (P/D = 48)••• as above + 0.90 M NaCl.

All in 0.05 M phosphate buffercontaining 0.1 M NaCl (v =0.2 M;pH = 6.8). A = 423 nm.

6XC

ion intensity of H.TMPyP is markedly quenched and the peaks are also

shifted to longer wavelengths (669 and 730 nm) compared to those found for

poly[d(A-T)]. The reduced emission intensity in this case is mainly a

consequence of less absorption of the excitation due to shifting of the

absorption maximum of H.TMPyP from 423 to 445 nm, but it is however4+

also partly due to a reduced quantum yield of fluorescence for H.TMPyP

when bound to poly[d(G-C)]. Making the solution 1 M in NaCl reduces the

observed effects, but in each case, and especially for poly[d(A-T)], there4+

is clear evidence for binding of a substantial fraction of the H.TMPyP to

the polynucleotide even under these high salt conditions. (In the absence4+

of DNA, 1 M NaCl has no effect on the emission spectrum of H.TMPyP ).

176

at Laurentian U

niversity on Novem


Dow

nloaded from



W A V E L E N G T H nm.

Figure 6. Fluorescence spectraof ZnTMPyP -̂DNA complexes!

(A) 2.3 yM ZnTMPyP4+

2.1 yM ZnTMPyP4+ and 420yM CT-DNA (P/D = 200)••• as above + 0.93 M NaCl.

(B) 1.7 yM ZnTMPyP4+

1.5 yM ZnTHPyP4+ and 110yM poly[d(G-C)] (P/D = 73)••• as above + 0.86 M NaCl.

(C) 1.7 yM ZnTMPyP4+

1.5 uM ZnTMPyP4+ and 140yM poly[d(A-T)] (P/D = 93)••• as above + 0.88 M NaCl.

All in 0.05 M phosphate buffercontaining 0.1 M NaCl (y = 0.2 M;pH = 6.8). X = 437 nm.* exc

4+The effect of polynucleotides on the fluorescence spectrum of ZnTMPyP

is shown in Figure 6. This spectrum (X at 632 nm, shoulder at ca.665 nm),

is hardly affected by polytd(G-C)], and an almost complete reversal of the

small reduction in intensity observed is caused by making the solution 1 M

in NaCl. With polytd(A-T)], however, two peaks are observed in the spec-

trum which are blue shifted (at 612 and 658 nm) compared to that of uncom-4+

plexed ZnTMPyP . 1 M NaCl causes a shifting of these peaks to 624 and4+

658 nm, but it is clear that much of the ZnTMPyP remains bound to the poly-

nucleotide. With CT-DNA pronounced splitting of the bands is also observed

the peaks being at 618 and 657 nm.

As addition of polynucleotide causes substantial changes in the inten-

177

at Laurentian U

niversity on Novem


Dow

nloaded from



sity of porphyrin fluorescence at certain wavelengths, fluorimetry may be

used to assay the extent of binding of the porphyrin and hence to calculate

binding constants. In optically dilute samples the intensity of emission

at a particular wavelength will be controlled by the absorption of the

excitation by the bound and unbound porphyrin (i.e. by their relative

concentrations and extinction coefficients at X ) and by their relative

fluorescence yield at the emission wavelength (X ). The intensity of

emission of the bound porphyrin may be estimated by extrapolation of the

data at high P/D values. From this and the intensity data at various P/D

ratios, the binding isotherm r/m versus r may be obtained (r = number of

moles of bound porphyrin per mole of base pairs of polynucleotide;

m = molar concentration of free porphyrin) (22). Preliminary experiments4+ -6

of this kind have been carried out for H.TMPyP (2.4 x 10 M) in

poly[d(G-C)] (X = 423 nm; X 668 ran), for H,TMPyP4+ (2.4 x 10~6 M)6XC GID *- i _*•

in poly[d(A-T)] (X = 423 nm; X = 655 nm) and for ZnTMPyP (1.6 x 10 M)

in poly[d(A-T)] (Xexc = 437 nm; Xem = 660 nm) all in 0.05 M lO^PO^/KjHPO^

0.1 M NaCl (p =0.2 M) buffer and with polynucleotides at concentrations so

that the P/D varied from 0 to 25 or greater. Analysis of these data

(corrected for the optically non-dilute conditions used) employing the

McGhee-von Hippel method (10,22,23) yields similar binding constants

(K ) of 1 - 2 x 106 M"1 for H,TMPyP4+ in poly[d(G-C)] , or in poly[d(A-T)]a P P 4 + •i

and for ZnTMPyP in poly[d(A-T)]. These K values may be compared toaPP

literature values obtained from absorption spectral data for H TMPyP in

polytd(G-C)] of 1.1 x 107 M" 1 (from Scatchard analysis) (1) and 7.7 x 105M~1

(from McGhee-von Hippel analysis) (10).Photolysis of DNA. Although DNA is not affected by low energy visible

light it has been reported that phosphodiester bonds may be photolysed in4+ 4+

the presence of porphyrins including H?TMPyP (4). The effect of ZnTMPyP

has not been reported. We have studied the effects of both compounds on

the photolysis of DNA by visible light by following the conversion of ccc

DNA to oc DNA to I DNA using pBR322 at a P/D ratio of ca. 50 (Figure 7).

A single break in the phosphodiester bonds of the sugar phosphate backbone

of pBR322 ccc DNA will convert it to oc DNA, and if two breaks occur in

opposite strands less than 5-10 base pairs from each other then oc DNA will

be converted to H DNA. The wavelength (436 nm) of the irradiation used was4+ 4+

close to the Soret maxima of the H.TMPyP and ZnTMPyP in DNA. No photo-

lysis of the DNA backbone occurred in the absence of porphyrin, and both

porphyrins sensitised photolysis. As shown in Figure 7 during the first178

at Laurentian U

niversity on Novem


Dow

nloaded from



Figure 7. H?TMPyP^+-induced phot-

olysis of plasmid DNA. pBR322 DNAirradiated for varying lengths oftime in the presence of H2TMPyP

4+

using light of 436 nm wavelengthat a P/D of ca.50. Lanes 1 to6, left to right, 0, 5, 10, 15,30 and 100 minutes exposure time.

15 minutes of irradiation the ccc DNA band disappeared with a corresponding

increase in the oc DNA band. After longer periods of irradiation a dis-

crete band corresponding to H DNA of unit size appeared,while some DNA

migrated faster than {, DNA in a continuous distribution, presumably smaller

DNA which had accumulated a large number of breaks.

Fiel et al. have noted that sodium azide (58 mM) , which is known to be a

quencher of singlet oxygen, completely inhibits the effects of three por-

phyrins (HPD, TCPP and TSPP) on the photolysis of DNA and partially inhib-4+

its the effect of H TMPyP (4). In our hands sodium azide (at 100 mM)4+ 4+

also partially blocked the effect of both H2TMPyP and ZnTMPyP (data notshown), confirming that singlet oxygen probably has a role in photolysis of

4+ 4+H TMPyP and that this is also probably the case for ZnTMPyP

DISCUSSION

The experiments reported here corroborate and extend the evidence that

both H TMPyP + and ZnTMPyP + interact with DNA. This is, for example,

demonstrated by their coelectrophoresis with DNA. The difference in the nat-

ure of this binding, however, is revealed by the experiment using topoisomer-4+ 4+

ase I. ZnTMPyP may unwind DNA to a very limited extent: H.TMPyP doeswith an efficiency similar to that of ethidium bromide, a known intercalator

4+(20). This is an explicit demonstration of intercalation for H.TMPyP , and

confirms the conclusions of earlier electrophoresis (2), viscosity (1,9) and4+

circular dichroism (1,6,7,8,10) experiments. If we assume that all H TMPyP

molecules are bound during topoisomerisation we may calculate the unwinding

angle per molecule of bound DNA. The measured change in supercoils per uM of4+

H^TMPyP added was 6 (Figure 3). This is equivalent to an unwinding angle

of 26° per residue, similar to that determined for ethidium bromide (24). No

179

at Laurentian U

niversity on Novem


Dow

nloaded from



4+diminution in the unwinding ability of H TMPyP is observed at the highest

concentrations tested (P/D = 32) so we conclude that not all intercalative/

unwinding sites have been filled at this ratio. This suggests there are more

than 300 intercalative sites per pBR322 molecule (4362 base pairs) for4+

H.TMPyP . As yet there are insufficient data on site specificity. The4+

binding constants for H TMPyP with poly(d(G-C)) and poly(d(A-T)) estimated

from fluorescence show that H TMPyP binds about equally to G-C and A-T4+

sites. In agreement with this the fluorescence spectrum of H TMPyP in

CT-DNA with 42% G-C is intermediate between that of poly(d(A-T)) and poly

(d(G-C)).4+

The effect of ZnTMPyP on topoisomerisation is much less than that of4+

H TMPyP (Figure 2) even though both compounds bind about equally tightly.4+

At the lowest concentration (0.4 pH ZnTMPyP ) equivalent to a P/D ratio of

200 there is a change of 2 - 3 negative supercoils, but this is the maximum

effect observed with no further change up to 3.2 ̂ M ZnTMPyP . It is poss-4+ .

lble that ZnTMPyP may unwind through interactions at a number of specific

sites which are saturated at P/D = 200 or more. This would be a particularly

interesting interpretation if true, and further experiments are being carried4+

out to test it. It is clear at this stage that the interaction of ZnTMPyP

with DNA differs quantitatively with respect to unwinding from that of

H2TMPyPA+.

As illustrated in Figures 5 and 6 interaction with DNA leads to sub-

stantial changes in the intensity and spectra of the fluorescence of the

porphyrins. Furthermore these changes depend on the base composition of

the polynucleotide. Upon interaction with poly(d(A-T)), for example, the

broad fluorescence band of H_TMPyP splits into two main bands with an

overall shift to higher energy and a very considerable increase in intensity.

As Kano et al. have recently pointed out the broad featureless fluorescence4+

band of H TMPyP is atypical of most monomeric porphyrins and these

authors have suggested that in aqueous solution this porphyrin may exist as

an aggregate even at concentrations as low as 10 M (26). Most other

evidence, however, points to H TMPyP being monomeric under these condit-

ions (27 and references therein). If H.TMPyP were to be present as a

dimer or other oligomer then the change of the fluorescence spectrum upon

interaction with DNA could be viewed, at least in part, as a consequence

of deaggregation. It should be noted however that enhancement of intensity

upon interaction of other dyes, such as acridines or phenanthridines,

with poly(d(A-T)) is well established (28,29) and this behaviour has been

180

at Laurentian U

niversity on Novem


Dow

nloaded from



ascribed largely to the removal of water from the solvation shell of the

dye. Indeed in the case of ethidium it has been argued that the excited

state is deactivated via proton transfer to water (30). It is striking

that the fluorescence spectrum of H TMPyP in poly(d(A-T)) very closely

resembles that in methanol (26) . This may be attributed to the H TMPyP

in this polynucleotide being in an environment similar to that of a polar

organic solvent and in particular to a reduction of water in its immediate

vicinity. By contrast in poly(d(G-C)) the emission quantum yield of4+ . . -1

H2TMPyP is somewhat reduced and the emission bands are at about 300 cm

to lower energy than in poly(d(A-T)). Quenching of the fluorescence of

dyes intercalated into poly(d(G-C)) is commonly found (28,29) and is consid-

ered to be a consequence of a redox interaction between the excited state

of the dye and guanine (29,31). A similar explanation is probably appro-

priate here as the singlet excited state of H?TMPyP is expected to be a

strong reductant. (E(P*/P ) may be calculated as +1.60 V using the data

in reference 27).

The fluorescence spectra of ZnTMPyP are markedly different in

poly(d(G-C)) and polyfd(A-T)). The spectrum in poly(d(G-C)) is very similar

to that in buffer, although there is a small (~20%) decrease in emission

intensity. This may be taken to indicate that any interaction of the por-

phyrin with the polynucleotide does not appreciably perturb its energy

levels. With poly (d(A-T)), however, a very marked change in the emission

spectrum is observed. This modification of the fluorescence may be con-

trasted with the rather slight changes observed in the absorption spectrum

of this porphyrin upon addition of the polynucleotide. The splitting of

the fluorescence into two bands and the overall shifting to higher energy

indicates that there must be a substantial modification of the environment

of the porphyrin molecule upon interaction. That this is not merely a weak

electrostatic binding is supported by the fact that making the solution 1 M

in NaCl only partly reduces the binding (whereas the spectrum of ZnTMPyP

in a solution containing poly(d(G-C)) and 1 M NaCl is closely similar to

that of the porphyrin by itself in buffer). Interestingly, unlike the

behaviour observed for TAJTMPy? a splitting of the broadened fluorescence

band is not observed in methanol indicating that merely removing most of the

water does not appreciably alter this compound's photophysical properties.

Indeed we have been unable to find a report of conditions in which ZnTMPyP

exhibits the type of spectrum found upon interaction with poly(d(A-T)). One

possibility is that when bound to the polynucleotide the molecule is held

181

at Laurentian U

niversity on Novem


Dow

nloaded from



more rigidly, leading to a sharpening of its features. This matter requires

further investigation.

In summary it may be seen from the absorption and fluorescence spectra

that poly(d(G-C)) strongly perturbs the electronic properties of H TMPyP4+

but has only a minor influence on those of ZnTMPyP . This is consistent

with intercalation at G-C base pairs occurring with the free base prophyrin4+

but not with its zinc complex. It is probable that ZnTMPyP is only bound

electrostatically to the outside of poly (d(G-C)). This differing behaviour

is probably a consequence of hindrance due to the axial ligand attached to

the zinc, as has been discussed by Pasternack et al. (10). On the other

hand, there are obvious and apparently similar changes in the emission

spectra of both H.TMPyP and ZnTMPyP when they interact with poly(d(A-T)).

The effect of salt shows that this binding cannot be merely electrostatic

and there is almost certainly substantial interaction with the bases. As the

topoisomerase experiment indicates that little unwinding occurs for the4+

ZnTMPyP , it is likely that this interaction with poly(d(A-T)) is non-

inter calative for this compound.

The conversion of pBR322 from its closed circular to open circular

form upon visible light irradiation in the presence of H^TMPyP or ZnTMPyP

indicates that this process causes single strand nicking. That both porphy-

rins show similar reactivity suggests that intercalation is not a pre-

requisite for photocleavage, as has already been pointed out by Fiel and co-4+ 4+

workers for H.TMPyP and other porphyrins (4). With both H.TMPyP and4+

ZnTMPyP the photosensitised cleavage is partially inhibited by 0.1 M

sodium azide, a known scavenger for singlet oxygen. The singlet oxygen is

presumably formed by reaction of the triplet state of the porphyrin with

ground state oxygen. In this connection it may be noted that the yield of4+

the triplet state is substantially reduced when H.TMPyP intercalates into

DNA (32), implying that the yield of 0 should also be lower in this system.

It is interesting to speculate whether the base specificity of intercalation,

the influence of the bases on the photophysical properties of the porphyrins

and the effects of salt or quenchers on the photochemical reaction may be

used to induce base selective cleavage reactions. These matters are under

investigation at present.

ACKNOWLEDGEMENT S

We thank B.P. Venture Research, Guinness & Co. Ltd., and the Investment

Bank of Ireland for their support of part of this research, Trinity College

Dublin, the Department of Education and Louth County Council for student-

182

at Laurentian U

niversity on Novem


Dow

nloaded from



ships for MJM and COhU, and the Radiation Laboratory University of Notre Dame

for a visiting fellowship for MJM during which some of this work and related

studies were carried out.

Abbreviations:

P/D : Molar concentration of UNA phosphate to dye (e.g. porphyrin).

ccc : covalently closed circular. Si : linear. oc : Open circular.

H TMPyP4+ and ZnTMPyP + : see Figure 1. CT-DNA : Calf thymus DNA.

DTT = dithiothreitol.

REFERENCES

1. Fiel, R.J., Howard, J.C., Mark, E.H. and Datta-Gupta, N. (1979)Nucleic Acids Res. 6, 3093-3118.

2. Fiel, R.J. and Munson, B.R. (1980) Nucleic Acids Res. 8, 2835-2842.3. Musser, D.A., Datta-Gupta, N. and Fiel, R.J. (1980) Biochem. Biophys.

Res. Commun. 97, 918-925.4. Fiel, R.J., Datta-Gupta, N., Mark, E.H. and Howard, J.C. (1981) Cancer

Res. 41, 3543-3545.5. Fiel, R.J., Beerman, T.A., Mark, E.H. and Datta-Gupta, N. (1982)

Biochem. Biophys. Res. Commun. 107, 1067-1074.6. Carvlin, M.J., Datta-Gupta, N. and Fiel, R.J. (1982) Biochem. Biophys.

Res. Commun. 108, 66-73.7. Carvlin, M.J. and Fiel, R.J. (1983) Nucleic Acids Res. 11, 6121-6139.8. Carvlin, M.J., Mark, E., Fiel, R.J. and Howard, J.C. (1983) Nucleic

Acids Res. 11, 6141-6154.9. Banville, D.L., Marzilli, L.G. and Wilson, W.D. (1983) Biochem. Biophys

Res. Commun. 113, 148-154.10. Pasternack, R.F., Gibbs, E.J. and Villafranca, J.J. (1983) Biochemistry

22, 2406-2414.11. Pasternack, R.F., Gibbs, E.J. and Villafranca, J.J. (1983) Biochemistry

22, 5409-5417.12. (a) Gouterman, M. (1978) in: The Porphyrins. 3, 1-65, ed. by Dolphin,

D., Academic Press, New York.(b) Darwent, J.R., Douglas, P., Harriman, A., Porter, G. and Richoux,M.C. (1982) Coord. Chem. Rev. 44, 83-126.

13. Little, R.G., Anton, J.A., Loach, P.A. and Ibers, J.A. (1975) J. Hetero-cyc. Chem. 12, 343-349.

14. Pasternack,R.F., Huber, P.R., Boyd, P., Engasser, G., Francesconi, L. ,Gibbs, E., Fasella, P., Cerio, V.G. and Hinds, L. (1972) J. Am. Chem.Soc. 94, 4511-4517.

15. Quimby, D.J. and Longo, F.R. (1975) J. Am. Chem. Soc. 97, 5111-5117.16. Kalyanasundaram, K. and Neumann-Spallart, M. (1982) J. Phys. Chem.

86, 5163-5169.17. Harriman, A., Porter, G. and Richoux, M.C. (1982) J. Chem. Soc.

Faraday Trans. II, 78, 1955-1970.18. Neta, P. (1981) J. Phys. Chem. 85, 3678-3684.19. Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Molecular Cloning,

a Laboratory Manual, Cold Spring Harbour Laboratory.20. Keller, W. (1975) Proc. Natl. Acad. Sci. USA 72, 4876-4880.21. Ishii, H., Satoh, K., Satoh, Y. and Koh, H. (1982) Talanta 29, 545-550.22. Dougherty, G. and Pigram, W.J. (1982) C.R.C. Crit. Rev. Biochem.

103-131.

183

at Laurentian U

niversity on Novem


Dow

nloaded from



23. McGhee, J.D. and von Hippel, P.H. (1974) J. Mol. Biol. 86, 469-489.24. Wang, J.C. (1974) J. Mol. Biol. 89, 783-801.25. (a) Dougherty, G. and Waring, M.J. (1982) Biophys. Chem. 15, 27.

(b) Dougherty, G. (1982) Comments Mol. Cell Biophys. 1, 337-354.26. Kano, K. , Miyake, T Yomoto, K., Sato, T., Ogawa, T. and Hashimoto, S.

(1983) Chem. Lett. 1867-1870.27. Kalyanasundaram, K. and Neumann-Spallart, M. (1982) J. Phys. Chem. 86,

5163-5169.28. Schreiber, J.P. and Daune, M.P. (1974) J. Mol. Biol. 83, 487-501.29. LBber, G. (1981) J. Lumin. 22, 221.265.30. Olmsted, J. and Kearns, D.R. (1977) Biochemistry 16, 3647-3654.31. Lober, G. and Kittler, L. (1973) Stud. Biophys. (Berlin) 73, 25-30.32. Murphy, M.J. (1984) Unpublished results.

184

at Laurentian U

niversity on Novem


Dow

nloaded from


a comparative study of the interaction of 5,10,15,20-tetrakis (n-methylpyridinium-4-yl)porphyrin and...

Documents